MVHM

MVHM数据集是由加利福尼亚大学欧文分校等机构创建的大型多视角手部网格数据集，旨在为3D手部姿态估计提供精确的网格和关节标注。该数据集包含320,000张分辨率为256×256的合成图像，每张图像都附带完整的手部关节和网格标注。创建过程中，研究者利用了TurboSquid提供的高质量手部模型，并通过Blender软件从不同角度渲染图像。MVHM数据集的应用领域广泛，包括人机交互、虚拟现实、增强现实等，旨在解决单视角3D手部姿态估计中的深度模糊问题。

The MVHM dataset is a large-scale multi-view hand mesh dataset developed by the University of California, Irvine and other institutions, aiming to provide accurate mesh and joint annotations for 3D hand pose estimation. This dataset contains 320,000 synthetic images with a resolution of 256×256, each accompanied by complete hand joint and mesh annotations. During its development, researchers utilized high-quality hand models provided by TurboSquid and rendered images from various perspectives using Blender software. The MVHM dataset has a wide range of application scenarios, including human-computer interaction, virtual reality, augmented reality and more, and is designed to address the depth ambiguity problem in single-view 3D hand pose estimation.